From heiner@geeni.bio.vtt.fi Sat Jul 31 12:07:42 1993 Date: Sat, 31 Jul 93 12:07:42 GMT From: heiner@bio.vtt.fi Message-Id: <9307311207.AA29575@geeni.bio.vtt.fi> To: chemistry@ccl.net Subject: GROMOS Mr. Breitkopf, in reply to your question about (graphical) input for the GROMOS package: what graphical, and what input. First off all, I've been working with Gromos for 7 years now, but I never encountered a original version that asked for graphical input. There is, though, a commercial variant from BioStructure in Strassbourgh who made a graphical interface to GROMOS, but in Groningen we never used it. Then, what input. Do you need input for the MD/EM section, or are you looking for building blocks that are used in the topology generator progmt/prormt. In the latter case you are more or less free to make your own, with force constants calculated from other packages or from literature, but you always have to evaluate the scientific worth of these parameters as the GROMOS ones are developed for amino acids/sugars. Especially when you mix your parameters with Gromos one's you have to be carefull (but that holds for any package). Whenever you are using the BioStructure package, it is most easy to ask them (you paid for the package after all). In case you use the original package, you may contact BIOMOS BV, assuming you have a legal version. You also can contact me directly at the address below. Pending these questions, Andrepeter Heiner. (formerly the Groningen group) ----------------------------------------------------------------------------- | A.P. Heiner, Ph.D. | Phone : +358-0-456 5105 | | VTT/Biotechnical Laboratory | Telefax : +358-0-455 2028 | | P.O. Box 202 (Tietotie 2) | E-mail : heiner@bio.vtt.fi | | FIN-02151 Espoo | | | FINLAND | | ----------------------------------------------------------------------------- From mercie@cumc.cornell.edu Sat Jul 31 06:13:54 1993 Date: Sat, 31 Jul 1993 10:13:54 -0400 (EDT) From: Gustavo Mercier Subject: Re: Ti parameters/basis sets To: Rob Stanton Message-Id: On Fri, 30 Jul 1993, Rob Stanton wrote: > > Recently I have been trying to find in the literature what has been done to > date with quantum mechanical calculations involving Ti. Any references would > be greatly appreciated. In particular references to previously used HF or DF > basis sets (pseudopotentials would be a bonus) or even published semiemperical > parameters would be helpful. In general, I would really like to hear about the > present state of the art in Ti calculations. > > Robert V. Stanton > Hi! Take a look at the following recent paper: Relativistic compact effective potentials and efficient, shared exponent basis sets for the third-, fourth-, and fifth-row atoms WJ Stevens, M. Krauss, H. Basch, and P.G. Jasien Can. J. Chem. 70, 612 (1992) This paper introduces a set of pseudopotentials with an introduction that describes previous work. The included references should give you a start. There are some typos as indicated to me by Dr. Krauss: Page 617,the In (Indium) Potential Vp-f should read for Ajk -129.78594, not -129.87594 Page 620, the basis set for Fe 1sp the Cp term should be -0.007940, not -0.07940 4sp the asp term should be 0.0410, not 0.410 Page 626, the Be Core polarizability Eval (calcd with core a) should read -1.01214, not -0.01214 If you wish full basis sets check handbook of gaussian basis sets, Physical sciences data 24 ed. r. poirier, r. kari, and i. g. csizmadia by Elsevier 1985 also check the work or Roos and Veillard like Roos, Veillard, and Vinot Theoret. chim. Acta (Berl.) 20, 1-11, 1971 Gaussian Basis Sets for Molecular Wavefunctions Containing Third-Row Atoms good luck gus From stoutepf@lldmpc.dnet.dupont.com Sat Jul 31 06:49:36 1993 Date: Sat, 31 Jul 93 10:49:36 -0400 Message-Id: <9307311449.AA23678@esds01.es.dupont.com> From: stoutepf@lldmpc.dnet.dupont.com (Pieter Stouten, +1-302-695-3515) To: "CHEMISTRY@ccl.net"@esds01.dnet.dupont.com Subject: RE: GROMOS On Fri, 30 Jul 1993 15:47, Mr. Breitkopf wrote: >I have a few questions concerning with the GROMOS-package. >I'll try to calculate 1,2-diols. >Since there are only some examples for the input of amino acids >I have problems to create input files and parameter files for not >predefined molecules. >Who has experience with such input files for a version without graphic >interface? > One respondent lectured on "scientific worth" and proved that reading is an art. Obviously, Mr. Breitkopf is aware of BioGromos, thinks that that is what most GROMOS users use, and informs us he does not have it. Also, the mere suggestion that Mr. Breitkopf might have an illegal version is appalling. Back to Mr. Breitkopf's original question, there is a program called PROMTI (+PROgram to Make Topology Input). Given atoms and connectivities, the program makes a formatted topology input file. Whenever it cannot resolve forcefield parameters, it indicates it. Also, it does not know about chirality and, therefore, leaves some improper torsions blank. It is up to the user to fill in the blanks. The major advantage of the program is that it defines the entire topology, so the user will make no errors by doing it manually, only by putting in the actual values :-). PROMTI can be obtained (free of charge as far as I know) from its author Bouke van Eijck (vaneyck@hutruu54.bitnet). Please be aware that mixing topologies produced by PROGMT and PROMTI (e.g. when studying enzyme/inhibitor systems) requires lots of manual intervention. For more GROMOS-related information, one may also consider subscribing to the GROMOS mailing list. Just send mail to gromos-request@igc.ethz.ch. The body of the mail should contain a line such as: please subscribe breitkopf@chemie.uni-halle.d400.de Hope this helps. Cheers, Pieter. Pieter Stouten, Senior Research Scientist Computer Aided Drug Design Group The DuPont Merck Pharmaceutical Company P.O. Box 80353, Wilmington, DE 19880-0353 Phone: +1 (302) 695 3515 Fax: +1 (302) 695 2813 ARA/Fax: +1 (302) 695 4324 E-mail: stoutepf@chemsci1.es.dupont.com E-mail: stoutepf@lldmpc.dnet.dupont.com From d3f012@pellucidar.pnl.gov Sat Jul 31 00:28:22 1993 Date: Sat, 31 Jul 93 07:28:22 -0700 From: d3f012@pellucidar.pnl.gov Subject: Ti Calculations To: chemistry@ccl.net Message-Id: <9307311428.AA21225@pellucidar.pnl.gov> > present state of the art in Ti calculations. > > Robert V. Stanton > Recently I have been trying to find in the literature what has been done to > date with quantum mechanical calculations involving Ti. Any references would > be greatly appreciated. In particular references to previously used HF or DF > basis sets (pseudopotentials would be a bonus) or even published semiemperical > parameters would be helpful. In general, I would really like to hear about th e > present state of the art in Ti calculations. > > Robert V. Stanton > Recently I have been trying to find in the literature what has been done to > date with quantum mechanical calculations involving Ti. Any references would > be greatly appreciated. In particular references to previously used HF or DF > basis sets (pseudopotentials would be a bonus) or even published semiemperical > parameters would be helpful. In general, I would really like to hear about th e > present state of the art in Ti calculations. > > Robert V. Stanton (Interrupt -- one more to kill letter) pellucidar 2% !! Mail chemistry@ccl.net Subject: Ti calculations > Recently I have been trying to find in the literature what has been done to > date with quantum mechanical calculations involving Ti. Any references would > be greatly appreciated. In particular references to previously used HF or DF > basis sets (pseudopotentials would be a bonus) or even published semiemperical > parameters would be helpful. In general, I would really like to hear about th e > present state of the art in Ti calculations. > > Robert V. Stanton There are INDO1 parameters for Ti published by the Zerner group. The reference is: "Calculated Spectra of Hydrated Ions of the First Transition-Metal Series" Wayne P. Anderson, W. Daniel Edwards, and Michael C. Zerner J. Inorg. Chem. 25, (1986), p. 2728. This particular paper does not give any Ti results. However it does explain the way these metals are parameterized and, of course, gives the parameter values as well as some examples using V,Cr,Mn, Fe,Co,and Ni. (caveat: Table I mistakenly gives units as 1000 cm-1. Actually, F0SS, F0SD, and F0DD are in ev, and the other params in this table are in 1000cm-1) ZINDO, Argus, and HyperChem's INDO support these atoms. Hope this helps. Mark ************************************************************************** Mark A. Thompson Sr. Research Scientist email: d3f012@pnlg.pnl.gov Molecular Science Research Center FAX : 509-375-6631 Pacific Northwest Laboratory voice: 509-375-6734 PO Box 999, Mail Stop K1-90 Richland, WA. 99352 Argus available via anonymous ftp from pnlg.pnl.gov (130.20.64.11) (in the argus directory). Download the README file first. Disclaimer: The views expressed in this message are solely my own and do not represent Battelle Memorial Institute, Pacific Northwest Laboratory, or any of its clients. ************************************************************************** From bbesler@ouchem.chem.oakland.edu Sat Jul 31 08:01:56 1993 Date: Sat, 31 Jul 93 12:01:56 -0400 From: bbesler@ouchem.chem.oakland.edu (Brent H. Besler) Message-Id: <9307311601.AA06440@ouchem.chem.oakland.edu> To: chemistry@ccl.net Subject: Gaussian 92 strange MP4 triples problems Newsgroups: comp.unix.aix Subject: Very strange problem with G92C Summary: Followup-To: Distribution: world Organization: Oakland University, Rochester MI. Keywords: I have run the computational chemistry application Gaussian 92 Rev C on an RS/6000 530 with 64 Mbytes of memory under AIX 3.2.3 and Fotran 2.1, 2.2, and 2.3. and also under AIX 3.15 with all of the above compilers. On a Cray C90, YMP, Silicon Graphics and IBM RS6000 340, 320H, another 530, 560, and 530H. All of these machines give me consistent results out to 9 decimal places(in the MP4 triples for those familiar with the application). The results on the errant 530 gives me different results in the 3rd to 8th decimal place(always different though). About 20% of the time with a small amount (16 Mbytes) of virtual memory used by the code the results will agree with the other machines. I have reformatted all of the disks of the 530 and installed only minimal AIX, the Fortran compiler and the application(Gaussian 92). IBM brought out two new memory boards which I tried with no luck. IBM's diags indicate no problem in memory, CPU or FPU. I suspect that either the FPU or a virtual memory management chip on the CPU board is bad. This RS600/530 was delivered in Sept. 1991, and was manufactured in May 1991. I think that this may be due to a deffective component on the CPU board due to the Myriad combinations of Fortran, C, and AIX combinations we have tried on the 530. Any opinions on this? From modesto@doctor.chem.yale.edu Sat Jul 31 12:19:46 1993 Date: Sat, 31 Jul 93 16:19:46 EDT From: Modesto Orozco Message-Id: <9307312019.AA01714@doctor.chem.yale.edu> To: CHEMISTRY@ccl.net Subject: charges for charged To K.Bryson and many other guys using semiempirical ESP charges. First: Never forget that when you use semiempirical charges instead of high level ab initio values you take your risks. The quality of the charges is done by the RMS of the fitting and also by their ability to reproduce other parameters like dipole (the SCF and if available the experimental), interaction energy, and others, if all these test are fine, the charges are fine. You must be aware that the ESP charges are only a set of scalar charges fitted to reproduce the expectation value of the r-1 operator in a lot of points out of the van der Waals radii. That's all, the ESP charges as a whole reproduce the MEP (the expectation value of the r-1 operator) well, but you have to use them knowing that they are not an exact magnitude (none charge is an expectation value!). Furthermore, they have a value as a whole, not as individual charges. F or instance, it has much more sense to look at the charges of CH3 as a whole tha n to pay too much attention to the individual charges of H and C. A lot of effor t is actually focussed in the development of other set of charges more reliable at the atomic level, but this is other history,... A note of hystoric justice: There are two strategies to obtain charges from semiempirical calculations: Deorthogonalization: Suggested simoultaneously by P.Kollman-K.Merz and our group (see vol 11 1990 of JCC). Pure NDDO suggested by Ferenczy, Reynolds and Richards (see the same volumen of the same journal). Most of the people use the first strategy (which I enjoyed very much), but the second one is also very good, and in my opinion merits much more attention. When your standard DO/semiempirical charges do not work well, it might be a good moment to prove the NDDO ones. Finally: If you use DP/semiempirical charges you might scale the charges (see Merz's or our papers), but ONLY if you deal with neutral molecules. If you consider a charged molecules you must not scale the charges. A possibility was to scale directly the MEP and compute the charges for the ions from this scaled MEP, whic h will mantein the total charge of the molecule. In a paper which will appear in few months in J.Comp.Aided Mol.Des., we demonstrate that the scaling factor ab initioMEP/semiempiricalMEP is 1.00 for cations and anions. That is you do not need to scale anything when compute semiempirical charges for cations and anions. The question of the quality is allways a debatable topic. From our results it seems that they are not bad, since the leading term in the determination of the MEP is the molecular monopole. NEVERTHELESS, remember: you take your risks. K.Merz Crys Reynolds, and ourselves have published several papers on the topic. Since a do not have the references in front of me, I have to suggest a quick reading to J.Comp.Chem. 14, 799 (1993), it contains most (at least several) of the major references in the field. Good luck!! Dr. Modesto Orozco Department of Chemistry Yale University Modesto@doctor.chem.yale.edu (203)432-6288